In a Long Term Evolution (LTE) system, a conventional antenna array has a fixed downtilt in a vertical direction, that is, provides a fixed beam for each user equipment in a cell in a vertical direction. A three-dimensional (3D) beamforming (BF) technology is introduced to improve a cell-edge user throughput rate and a cell average throughput rate. In the 3D beamforming technology, based on an active antenna system (AAS), beams with different downtilts can be generated for each user equipment in a vertical direction according to a location of the user equipment, so that beamforming is performed in both a horizontal direction and the vertical direction. To support the 3D beamforming technology, corresponding channel state information needs to be fed back. The channel state information includes a channel quality indicator (CQI), a precoding matrix indicator (PMI), and a rank indication (RI).
The channel state information is usually obtained by user equipment by means of channel estimation. In a conventional channel estimation process, user equipment measures a reference signal (RS), which is also referred to as a pilot signal, such as a channel state information-reference signal (CSI-RS), then obtains channel state information in a horizontal dimension by means of estimation, and reports the channel state information to a base station. However, for the 3D beamforming technology, in addition to performing channel estimation on a channel in a horizontal dimension, the user equipment needs to perform channel estimation on the channel in a vertical dimension. Obviously, a conventional channel estimation method that supports channel estimation in only a horizontal dimension cannot be applicable to the 3D beamforming technology. Therefore, how to propose a channel state information feedback method becomes a problem to be urgently resolved in the industry.